Automated Medicinal Plant Recognition Using CNN | IJCT Volume 13 – Issue 3 | IJCT-V13I3P12

Medicinal plants have been an integral part of traditional healthcare systems such as Ayurveda, Siddha, and Unani for centuries. Even in modern medicine, a significant percentage of pharmaceutical drugs are derived from plant-based compounds. Accurate identification of medicinal plants is therefore essential for ensuring safe and effective treatment. However, manual identification requires botanical expertise and is often affected by human error, environmental variations, and morphological similarities among plant species. This research proposes an intelligent and automated medicinal plant recognition system using Convolutional Neural Networks (CNN) optimized with Particle Swarm Optimization (PSO). The proposed system performs multi-functional tasks including plant species classification, plant disease detection, and integration of Ayurvedic medicinal knowledge. The CNN model is designed to extract deep features from plant leaf images, while PSO is applied to optimize hyperparameters for improving classification accuracy and convergence speed. The system also incorporates a disease detection module capable of identifying fungal infections and physical damages in leaves. A structured Ayurvedic database provides medicinal properties, therapeutic uses, and traditional remedies associated with identified plant species. Furthermore, a user-friendly graphical interface enables easy interaction for farmers, researchers, and practitioners. Experimental results demonstrate that the PSO-optimized CNN achieves higher accuracy compared to traditional CNN models. The proposed framework supports sustainable agriculture, medicinal plant conservation, and accessible healthcare by combining artificial intelligence with traditional knowledge systems.